In a wavelength division multiplexing optical transmission system, various information channels are encoded into light at different wavelengths, which is combined using a multiplexor. The combined light is transmitted through an optical fiber and, or an optical fiber network to a receiver end of the optical fiber. At the receiver end, the signal is separated, or demultiplexed, back into the individual optical channels through a de-multiplexor, whereby each optical channel can be detected by an optical detector such as a photodiode, and the information can be reconstructed, channel by channel.
While propagating through the optical fiber, light tends to lose power due to the losses related to the physics of how the light interacts with the optical fiber. Yet some minimal level of optical channel power is required at the receiver end in order to decode information encoded in the optical channel. In order to boost the optical signal propagating in the optical fiber, optical amplifiers are deployed at multiple locations, known as nodes, along the transmission link. The optical amplifiers extend the maximum possible length of the link, in some instances, from a few hundred kilometers to several thousand kilometers, whereby after each fiber span, the optical signal is amplified to power levels close to the original levels of the transmitter. Unfortunately, during the amplification process some amount of noise is introduced into the optical signal which effectively limits the amount of optical amplifiers a transmission link can have.
Modern optical communication systems employ erbium doped fiber amplifiers (EDFAs), Raman Amplifiers (RAs) and hybrid EDFA-RAs as means to boost the optical signal power and thus to extend the communication system reach. Nowadays, optical communication systems have become more agile and reconfigurable. Reconfiguration of the optical communication system leads to variation of the signal load at the input of the amplifier. At the same time, the goal of the amplifier is to provide constant gain, which should not depend on the power or wavelength loading condition; otherwise, some channels will not have sufficient power and signal-to-noise level at the receiver end, resulting in information being lost.
The control electronics of EDFAs partially solves the problem of the variable signal load. More particularly, the total optical power at the input and at the output of the amplifier is measured, and the average optical signal gain of the amplifier is calculated. The amplifier control electronic circuitry adjusts the amplifier's pump powers through a feedback loop in such a way that the measured optical gain equals to the desired or “set” optical gain and is not varied significantly in time.
However, it is desired not only to have average gain of the amplifier to be constant, but also to have the gain of the individual channel constant and independent from the other channels' presence or absence, that is, independent from the channel load. At the same time, due to the spectroscopy of the erbium doped fiber, namely due to the spectral hole burning (SHB) effect, the gain shape of EDFA does depend on the input load. Hence even if the average gain of an EDFA is held constant, the gain of the individual channels will vary, leading to undesirable effects, such as increased bit error rate of the transmission system.
One way in which to address the problem is to check the channel powers at a location in the transmission system, using an optical channel monitor (OCM). The collected information is then used by the system control circuitry to adjust a dynamic gain equalizer (DGE) in the transmission link in such a way that the transmitted spectrum is flattened. The OCM and DGE need not necessarily be at a same location in the system. The advantage of this approach that it compensates for all gain change inducing impairments of the system, such as stimulated Raman scattering (SRS) induced tilt, not only EDFA SHB.
However the above approach has several disadvantages. First, because the DGE and OCM are expensive components, they are not generally installed at each amplifier node, thus they compensate several amplifiers at once, which is not optimal. Second, both OCM and DGE are comparatively slow devices, and thus the correction usually takes a few seconds. This is undesirable for agile communication systems where a typical requirement for the adjustment for a transient event such as a change of the channel load is on the order of 100 μs, which is 10,000 times shorter than for a DGE/OCM approach.
To address the disadvantage of this compensating technique it has been suggested by Zhou et al. in an article entitled “Fast control of inter-channel SRS and residual EDFA transients using a multiple-wavelength forward-pumped discrete Raman amplifier”, OMN4, OFC 2007, which is incorporated herein by reference, to measure channel powers of a limited number of channels that are located at specific wavelengths, 1528.6 nm, 1544.4 nm, and 1559.6 nm in the published example. Subsequently, the Raman pump powers of the Raman amplifier are adjusted using linear feed-forward control. The work is based on RAs having 3 different wavelengths of Raman pumps. Again, similar to the aforementioned DGE/OCM approach, this compensates not only EDFA SHB, but SRS tilt as well.
The main disadvantage of this technique is the requirement of the constant presence of those three channels the power of which is constantly monitored. This is a very limiting requirement for modern agile communication systems. Another potential disadvantage is the requirement to have three additional detectors. Finally, relatively good SHB compensation is possible only in the presence of three Raman pumps—the reduction of number of pumps will lead to the reduction of the amount of compensation.
Further, in U.S. Pat. No. 7,359,112 by Nishihara et al. which is incorporated herein by reference, a control apparatus is described which adjusts the gain of an EDFA based on an amount of wavelengths which is calculated on the basis of optical power measured in two or three separate spectral bands by dedicated photodetectors. One disadvantage of this approach is that only one control parameter, specifically the EDFA gain, is adjusted which limits the degree to which both the SHB and SRS can be compensated. Another disadvantage stems from the fact that certain load change patterns, for example the patterns which leave the total optical power measured in a single spectral band unchanged, will not be detected by the apparatus of Nishihara et al. and therefore will not be compensated for by said apparatus.
It is an object of the present invention to provide an apparatus and method for flattening a gain profile of an optical amplifier suitable for suppression of sub-millisecond scale transient variations of gain caused by changes in the amplifier load which would not require dedicated spectral channels in order to monitor the gain profile. This invention extends the technique that was suggested by Bolshtyansky et al. in an article entitled “Dynamic Compensation of Raman Tilt in a Fiber Link by EDFA during Transient Events”, JThA15, OFC 2007, where instead of measuring the actual gain change, the device measures some property of the transmitted signal, and adjusts the gain profile based on the measured property of the signal.